Milky Way galaxy's central black hole is spinning at maximum speed and pointing at Earth

A new analysis argues that the Milky Way’s central black hole, Sagittarius A*, spins at maximum velocity with its axis aimed almost directly at Earth.

The result emerges from a machine learning read of the Event Horizon Telescope data that pulls out subtle signals from ordinary methods left on the table.

Those signals sit on top of a now famous foundation. In 2017, the Event Horizon Telescope released the first image of a black hole, the ring around M87*. 

How a neural network read the data

The study was led by Michael Janssen of Radboud University (RU) with collaborators across Europe and the United States, and it used computing tools and expertise from the University of Southern California’s Information Sciences Institute (USC ISI). 

At the heart of the analysis is a Bayesian neural network, a type of model that returns both a best estimate and an uncertainty.

The team trained it on an enormous library of synthetic observations that mimic the telescope’s raw measurements, not just pictures on a grid.

Instead of feeding heavily averaged images to the algorithm, the researchers taught it the full polarization visibilities in time and frequency, the native language of very long baseline interferometry (VLBI).

That choice preserves information and lets the model find stable patterns even when some measurements are noisy.

Why spin and tilt matter

Black hole spin is a number between 0 and 1 that captures how fast the object rotates. A higher value means more rotational energy available to shape the flow of plasma and, in some systems, to help power a jet.

Orientation matters, too. When the spin axis aligns with our line of sight, the telescope records a distinct signal profile compared to an edge on orientation, with the strongest differences appearing in polarized light.

What the data say about the Milky Way

The network’s posterior for Sagittarius A* favors a high spin, about 0.8 to 0.9, and a small tilt with the axis close to our line of sight.

This is consistent with an axis angle near 162 degrees in one family of models and about 29 degrees in another.

That combination is the technical way of saying the black hole’s axis points almost directly toward Earth.

The same fit suggests the light near the hole is dominated by extremely hot electrons in the surrounding accretion disk, not by a jet.

In the team’s parameters, the preferred electron temperature ratio is around 14 for Sagittarius A*, a disk leaning value that matches the polarized signatures they measure.

Heavy lifting behind the scenes

Training required on the order of a million realistic, calibrated synthetic datasets, each passed through a forward model of the instrument and atmosphere.

The framework itself, called Zingularity, leans on TensorFlow Probability and Horovod to scale across modern accelerators.

The pipeline that produced those datasets mirrored the real reduction path, including calibration steps and error bootstrapping, and ran across an ecosystem that kept the neural network honest about the ways real measurements can be corrupted.

“The ability to scale up to the millions of synthetic data files required to train the model is an impressive achievement,” said Chi kwan Chan of the University of Arizona, who works on both the simulations and the data systems. 

What about M87

The same approach revisited M87*, the first black hole ever imaged, and found a different character.

The preferred solutions point to a fast, retrograde spin, roughly between -0.5 and -0.94 with respect to the flow of gas spiraling inward, and to strong synchrotron emission from the jet.

That retrograde result means the black hole rotates opposite to the direction of the infalling plasma. The team suggests a past merger could have flipped the alignment and left behind the counter rotation they now infer.

What comes next

The network was trained on comprehensive simulations, but the authors are clear about model limits and the value of better data.

New stations, more bandwidth, and higher observing frequencies will tighten the constraints and reduce degeneracies in the fits.

A key addition on the horizon is the Africa Millimetre Telescope planned for Namibia, which will lengthen baselines into the Southern Hemisphere and sharpen the array’s view of Sagittarius A*.

As the array grows, rapid cadence and multi frequency observations will let models test whether the Milky Way’s black hole really is a high spin, face on system and map how magnetic fields feed energy into the flow. That is the path to distinguishing between look alike fits and turning hints into hard numbers.

The study is published in Astronomy & Astrophysics.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–